The current study aims to apply and compare the performance of six machine learning algorithms, including three basic classifiers: random forest (RF), gradient boosting decision tree (GBDT), and extreme gradient boosting (XGB), as well as their hybrid classifiers, using the logistic regression (LR) method (RF+LR, GBDT+LR, and XGB+LR), in order to map the landslide susceptibility of Zhangjiajie City, Hunan Province, China. First, a landslide inventory map was created with 206 historical landslide points and 412 non-landslide points, which was randomly divided into two datasets for model training (80%) and model testing (20%). Second, 15 landslide conditioning factors (i.e., altitude, slope, aspect, plane curvature, profile curvature, relief, roughness, rainfall, topographic wetness index (TWI), normalized difference vegetative index (NDVI), distance to roads, distance to rivers, land use/land cover (LULC), soil texture, and lithology) were initially selected to establish a landslide factor database. Thereafter, the multicollinearity test and information gain ratio (IGR) technique were applied to rank the importance of the factors. Subsequently, we used a series of metrics (e.g., accuracy, precision, recall, f-measure, area under the ROC (receiver operating characteristic) curve (AUC), kappa index, mean absolute error (MAE), and root mean square error (RMSE)) to evaluate the accuracy and performance of the six models. Based on the AUC values derived from the models, the GBDT+LR model with the highest AUC value (0.8168) was identified as the most efficient model for mapping landslide susceptibility, followed by the XGB+LR, XGB, RF+LR, GBDT, and RF models, which achieved AUC values of 0.8124, 0.8118, 0.8060, 0.7927, and 0.7883, respectively. The results from this study suggest that the stacking ensemble machine learning method is promising for use in landslide susceptibility mapping in the Zhangjiajie area and is capable of targeting the areas prone to landslides.
The optimization of groundwater conditioning factors (GCFs), the evaluation of groundwater potential (GWpot), the hydrogeological characterization of aquifer geoelectrical properties and borehole lithological information are of great significance in the complex decision-making processes of groundwater resource management (GRM). In this study, the regional GWpot of the Karak watershed in Northern Pakistan was first evaluated by means of the multi-influence factors (MIFs) model of optimized GCFs through geoprocessing tools in geographical information system (GIS). The distribution of petrophysical properties indicated by the measured resistivity fluctuations was then generated to locally verify the GWpot, and to analyze the hydrogeological and geoelectrical characteristics of aquifers. According to the weighted overlay analysis of MIFs, GWpot map was zoned into low, medium, high and very high areas, covering 9.7% (72.3 km2), 52.4% (1307.7 km2), 31.3% (913.4 km2), and 6.6% (44.8 km2) of the study area. The GWpot accuracy sequentially depends on the classification criteria, the mean rating score, and the weights assigned to GCFs. The most influential factors are geology, lineament density, and land use/land cover followed by drainage density, slope, soil type, rainfall, elevation, and groundwater level fluctuations. The receiver operating characteristic (ROC) curve, the confusion matrix, and Kappa (K) analysis show satisfactory and consistent results and expected performances (the area under the curve value 68%, confusion matrix 68%, Kappa (K) analysis 65%). The electrical resistivity tomography (ERT) and vertical electrical sounding (VES) data interpretations reveals five regional hydrological layers (i.e., coarse gravel and sand, silty sand mixed lithology, clayey sand/fine sand, fine sand/gravel, and clayey basement). The preliminary interpretation of ERT results highlights the complexity of the hydrogeological strata and reveals that GWpot is structurally and proximately constrained in the clayey sand and silicate aquifers (sandstone), which is of significance for the determination of drilling sites, expansion of drinking water supply and irrigation in the future. Moreover, quantifying the spatial distribution of aquifer hydrogeological characteristics (such as reflection coefficient, isopach, and resistivity mapping) based on Olayinka’s basic standards, indirectly and locally verify the performance of the MIF model and ultimately determine new locations for groundwater exploitation. The combined methods of regional GWpot mapping and hydrogeological characterization, through the geospatial MIFs model and aquifer geoelectrical interpretation, respectively, facilitate decision-makers for sustainable GRM not only in the Karak watershed but also in other similar areas worldwide.
In order to get a higher success rate and a small number of wells for development, the reservoir characterization was carried out using seismic, Petrophysical and rock physics data in the B-interval of Miano 9 and 10, Miano area, Lower Indus Basin, Pakistan. For this purpose, at first, the wells and shot points of seismic lines were located on a base map. Then, the horizons were marked on seismic sections and further interpreted on the Time Depth (TD) charts. Moreover, a synthetic seismogram relaying the seismic and well data was made. In addition, different Petrophysical parameters were calculated using well logs and the geological properties of rocks were related with the corresponding elastic and seismic properties in analyzing the rock physics. The study indicates that the area is deformed by horst and graben structure in subsurface and lateral extents of faults were marked by fault polygons. The higher hydrocarbon saturation, porosity and permiability as determined through well logs and the low elastic properties calculated through rock physics in B-interval of Miano 9 and 10 indicate a good potential zone of Hydrocarbon. The finding of this research work is useful to understand the subsurface geology and hydrocarbon potential of the area.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.